Super cool!

Looks like all the data is in a fairly decently scrapable format. Could one, in theory at least, add auto-identificaton to an SDR app?

Is there an (possibly ML) algorithm that says "take this single example picture and tell me if there is a match(Which could be scaled differently) in this other picture"?